*3.3. L. braziliensis HSP23- and HSP100-Null Mutant Phenotypes Resemble Those Described for Old World Leishmania*

For the phenotype analysis, we first attempted to create gene add-back parasites for both null mutants. In the *HSP23*–/– mutants, we introduced the *LbrHSP23* transgene for integration into the 18S SSU rRNA locus, using the pIRmcs3+ vector [53], or as episome, using the over expression plasmid pCL1S-*LbrHSP23*. To generate the *HSP100* add-back cell lines, the *HSP100*–/– mutants were transfected with the pIRmcs3+ vector harbouring *LbrHSP100* for genomic integration. Despite several attempts with different experimental conditions (data not shown), we could not generate any of the intended gene add-back cell lines. We suspect that the selection marker gene, coding for streptothricine N-acetyl transferase (SAT), was not stably expressed, possibly due to the known RNAi activity in *L. braziliensis*[9]. Ectopic gene expression from integrated and episomal transgenes is unpredictable in *L. braziliensis* (V.A., unpublished observations, and [59]).

We nevertheless proceeded to test the growth phenotypes of the *L. braziliensis HSP23*–/– and *HSP100*–/– null mutants under various *in vitro* growth conditions compared with the wild-type and with Cas9-expressing cells. Cell density on day 4 (stationary phase) was analysed and displayed as percentage of growth relative to the wild type (set at 100%). Under optimal *in vitro* growth conditions for promastigotes (25 ◦C, pH 7.4), the *L. braziliensis* PER005cl2 wild-type strain achieved a median 24.9-fold growth (2.49 × 10<sup>7</sup> cells/ml). Two *HSP23*–/– null mutants, *HSP23*–/– cl.2 and *HSP23*–/– cl.3, grew at rates similar to the wild type (median relative growth: 85.0% for *HSP23*–/– cl.2 and 93.4% for *HSP23*–/– cl.3; Fig. 4A). *HSP23*–/– cl.1 displayed a 20% elevated proliferation, similar to the Cas9-expressing cells. The *HSP100*-null mutants showed proliferation rates (median relative growth: 86.1% for *HSP100*–/– cl.1 and 81.8% for *HSP100*–/– cl.2) comparable to those of the wild type (Figure 4A). Therefore, we see no growth phenotype for *HSP23*–/– and *HSP100*–/– null mutants under optimal culture conditions. This is in keeping with earlier findings about the significance of HSP100 and HSP23 in the promastigote [47,56]. Stable Cas9 expression from the pTB007 episome increased the growth rate of *L. braziliensis* promastigotes at 25 ◦C (Figure S1C), leading to a higher cell density in late-log phase (day 3; *p* = 0.004, *U* test) and in stationary phase (day 4; *p* = 0.015, *U* test) compared to the wild-type parasites, likely reflecting a positive effect on cell proliferation, similar to previous observations [21].

Next, we repeated the analysis at 30 ◦C, the upper temperature limit for *L. braziliensis* growth *in vitro* [60]. Proliferation of the *L. braziliensis* PER005cl2 wild-type strain was slowed considerably at 30 ◦C, reaching a median of 4.9 × 10<sup>6</sup> cells/ml at day 4 (4.9-fold growth). The *L. braziliensis HSP23–*/*–* null mutants, particularly *HSP23*–/– cl.2 and *HSP23*–/– cl.3, were sensitive to the 30 ◦C cultivation temperature and did not proliferate (Figure 4B). This temperature-sensitive phenotype is in line with previous work with *L. donovani HSP23–*/*–* null mutants [47]. We also tested the cell integrity of the *L. braziliensis HSP23*–/– null mutants at 30 ◦C. As shown by immunofluorescence microscopy (Figure 4C), all three *L. braziliensis HSP23*-null mutants showed abnormally rounded, swollen and irregular shapes, and formed cell aggregates indicating cellular damage. These changes were not observed in the control cells, *L. braziliensis* wild type and Cas9-expressing cells, which presented as individual, well defined cells.

**Figure 4.** Phenotypic analyses of *L. braziliensis HSP23–*/*–* and *HSP100–*/*–* clones. For growth curves, promastigotes of WT, WT (Cas9), *HSP23*–/– clones, and *HSP100*–/– clones were seeded at a density of 1 × 10<sup>6</sup> parasites/mL into 5 ml of complete M199 medium and grown for 4 days. Cell density was measured on day 4 and is shown as a percentage of WT cell density (set at 100%). Parasites were grown at 25 ◦C (**A**) and 30 ◦C (**B**). The *HSP23*–/– clones incubated for 4 days at 30 ◦C were also stained with mouse anti-tubulin antibody (1/4000) and DAPI (1/50) (**C**). Images were taken on an EVOS FL Auto Cell Imaging System and processed using the ImageJ Software (https://fiji.sc). Scale bar: 10µm. Additional cultures were grown at 25 ◦C and pH 7.4 with the addition of 2% ethanol (**D**). The horizontal black lines in panels A, B, and D indicate the median of 6 biological samples from 3 separate experiments. Significance was tested using the Kruskal–Wallis test; \* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.001. (**E**) Primary mouse bone-marrow-derived macrophages were differentiated and infected with stationary-phase promastigotes of WT, WT [Cas9], *HSP23*–/– clones, and *HSP100*–/– clones at a MOI of 1:8 (macrophage-to-parasite ratio). After 4 h, free parasites were washed away and the infected macrophage cultures were further incubated at 34 ◦C under 5% CO<sup>2</sup> for 44 h. Genomic DNA from *Leishmania*-infected macrophages was isolated at 4.5 h and at 48 h post-infection, and parasite load was determined by TaqMan qPCR quantifying parasite *actin* gene DNA relative to host macrophage *actin* gene DNA. Shown is intracellular parasite survival [%] after 48 h, with the bar indicating the median of *n* = 5. Ratio-paired, one-sided Student's *t*-test: \* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.001 between data pairs. ns = not significant.

Conversely, the *L. braziliensis HSP100*–/– null mutants were fully viable and proliferating at 30 ◦C, even exhibiting a significant growth advantage over the wild type (Figure 4B). This temperature tolerance of the *L. braziliensis HSP100*–/– null mutants matches previous findings from phenotype analyses of *L. donovani HSP100*–/– null mutants [57], but contrasts with the phenotype of *L. major HSP100*–/– null mutants, which were hypersensitive at the upper limit of growth temperature [56]. Lastly, the Cas9-expressing cells grown at 30 ◦C also showed an elevated growth without reaching statistical significance (Figure 4B).

We next tested the *L. braziliensis HSP23*–/– and *HSP100*–/– null mutants for tolerance to sublethal ethanol concentrations, a trigger of the unfolded protein response, a stress signalling pathway of the endoplasmic reticulum (ER) that is related to the heat shock response [61,62]. Treatment with 2% ethanol caused growth reduction for all three *L. braziliensis HSP23*–/– null mutants (Figure 4D). This increased sensitivity of *L. braziliensis HSP23*–/– null mutants to a chemical stressor (i.e., ER stress-sensitive phenotype) is in agreement with previous work in *L. donovani HSP23*–/– mutants [47], further supporting the involvement of HSP23 in protecting *Leishmania* against protein misfolding stress. The *HSP100*–/– null mutants were not affected by exposure to 2% ethanol (Figure 4D). Again, the Cas9-expressing cells showed a slightly increased growth compared to the wild type (Figure 4D).

Lastly, we tested the ability of the wild type and mutant strains to survive inside macrophages. Primary mouse bone marrow-derived macrophages were differentiated and infected *in vitro* at a parasite to macrophage ratio of 8:1 using stationary-phase promastigotes. The parasite load was evaluated by qPCR [50] at 48 h post infection relative to the parasite load after 4.5 h of parasite internalisation.

The average percentage of surviving *L. braziliensis* PER005cl2 wild-type parasites within macrophages at 48 h post-infection was 52.6 ± 13.0% (Figure 4E). The loss of HSP100 had a significant impact on the intracellular survival of the two *L. braziliensis HSP100* null mutants. The effect was more pronounced for the whole-gene deletion mutant (*HSP100*–/– cl.2; mean survival ± SD: 23.4 ± 11.7%) than for the partial gene disruption (*HSP100*–/– cl.1; 32.0 ± 12.4%) (Figure 4E). The impaired ability of these *L. braziliensis HSP100*–/– null mutants for intracellular survival in *in vitro*-infected mouse macrophages was also documented for *L. major* and *L. donovani HSP100*-null mutants [56,57].

The ability to survive in macrophages was affected in only two *L. braziliensis HSP23*–/– mutants (*HSP23*–/– cl.2: mean survival ± SD: 21.2 ± 6.1%; *HSP23*–/– cl.3: 35.6 ± 12.5%)(Figure 4E), whereas the *HSP23*–/– cl.1 was able to survive intracellularly (54.1 ± 16.8%) at a rate similar to the wild-type parasites (Figure 4E). The reduced survival of *HSP23*–/– cl.2 and cl.3 matches the poor growth of these clones at elevated temperature and under ethanol stress (Figure 4B,D) and is in line with previous work performed with a *L. donovani HSP23*-null mutant [47].

In a first attempt to investigate possible genomic adaptations in the mutants as cause for varying phenotypes, we evaluated aneuploidy patterns. Using the NGS sequence reads from the WGS analysis and quantifying normalised sequence read densities for individual chromosomes in *L. braziliensis* WT cells, WT [Cas9] cells, three *HSP23*–/– mutant clones and two *HSP100*–/– mutant clones, we calculated chromosome ploidies (Figure S9A). Indeed, we found profound differences between *L. braziliensis HSP23*–/– mutants themselves and compared to the other parasite strains. *HSP23*–/– clone 1 is trisomic for chromosome 30 and shows intermediate somy (2.56) for chromosome 4. *HSP23*–/– clone 2 shows a marked increase of chromosome 2 ploidy (4.82). *HSP23*–/– clone 3 shows strong amplification (4.6) of chromosome 14, trisomies for chromosomes 18, 33 and 34, and a slight (2.39) increase for chromosome 4, which was also partly amplified in *HSP23*–/– clone 1. The strong increase of chromosome 2 sequence reads for *HSP23*–/– clone 2 is due to an apparent amplification of a ~20,000 bp region between positions 260,000 and 280,000 (Figure S9C). The amplified region contains mostly copies of a SLACS retrotransposon (LbrM.02.0550), and a possible context with the loss of HSP23 is not obvious.

All three *L. braziliensis HSP23*–/– clones, but also the Cas9-expressing strain were trisomic for chromosome 26, possibly causing the minor fitness gain observed for the Cas9 strain.
